Computer Crashing - Is my GPU responsible?

[_kain]

Starting a little bit over a month ago, I've been getting 2 different types of crashes (BSODS) and I really need help in figuring out whats causing them so I can try and fix the problem, or get new hardware.

(Crash 1)
The first type of crash is specific to the videogame Dota2. I notice loading screen times take alot longer than usual, and then anywhere from 2-10 minutes of playing the game, the video will hang. I can still hear ingame sounds/music/voices of my friends on teamspeak but the monitor will be frozen to whatever was onscreen at the time.

The only way I'm able to solve this crash is to fully remove my drivers using a program called 'Display Driver Uninstaller.exe' and download the newest drivers from Nvidia (although sometimes I believe there isn't even a new driver so I'm just re-installing the same one).

(Crash 2)
The second crash I get can happen at any time, most often whilst playing a game but even on desktop or browsing the web. When this happens the monitor just turns black (it still has power, but looks as if its in standby mode) and computer becomes unresponsive.

The only way I've managed to get the computer to boot again is to power it off, unplug everything and hold down the power button for 10sec and then reboot.

Crash 1 usually happens maybe once a week, Crash 2 started happening maybe once a week, but has been happening more and more often.

(Crash Logs)
After experiencing these crashes a few times I decided to download BlueScreenView to log the crashes. For some reason it doesnt log every one, but here is an image of the ones it has logged:

I believe the 'ntoskrnl.exe' crash is the one that happens in Dota2, and 'dxgkrnl.exe' is Crash 2.

From what I've researched these crashes can be caused by a faulty Mobo, GPU or bad RAM. I've used windows memtest tool to check RAM and it found no problems.

From what I've read bad video drivers can also cause these problems, but I've completely removed old drivers and installed new ones several times.

I'm hoping that someone can point me towards it being GPU or Mobo (or something else I've not thought of) as I don't have alot of money and can't afford to replace both.


Computer Specs:

• 
  Processor: i7 960 @3.2Ghz
  GPU: GTX 680
  RAM: 12288MB
  OS: Windows 7

 

[Endless8]

Hello _kain,
I see you did well with some research already. But let's start with few simple tests.
Download HWMonitor to check temps, prime95 to test CPU, UNrealHeaven, 3DMark or Furmark to test GPU and see what happens, no errors in MemTest is a good sign.
You can try removing drivers with DDU and install beta ones (if avaible), i had similar issue once and I fixed it like this, after the problem did not appeared with new drivers. You can also update DirectX.
Btw what is your PSU? and do you have the possibility to test your card in another PC or try another card (if you have one) in yours?
Cheers.

 

[_kain]

Hello _kain,
I see you did well with some research already. But let's start with few simple tests.
Download HWMonitor to check temps, prime95 to test CPU, UNrealHeaven, 3DMark or Furmark to test GPU and see what happens, no errors in MemTest is a good sign.
You can try removing drivers with DDU and install beta ones (if avaible), i had similar issue once and I fixed it like this, after the problem did not appeared with new drivers. You can also update DirectX.
Btw what is your PSU? and do you have the possibility to test your card in another PC or try another card (if you have one) in yours?
Cheers.

Hi Endless, thanks for the reply. I have OpenHardwareMonitor installed and on desktop temps seem fine (and the crash can happen when I'm just on the desktop and using a web browser) everything is about 30C except CPU which is about 35-37C.

I tried updating DirectX but already at the latest version. I will download prime95 and a GPU tester now and get back to you.

note: I cant remember the exact power of my PSU, but my system has been running fine for 1year+ (which is the last time I upgraded it when i bought my current PSU and GPU, and I made sure that I bought a PSU with more than enough power). Unfortunately I don't have another card/PC to test.


EDIT AFTER TESTING:

I ran Prime95 and CPU temperatures rose until they stabilized at about 61C - No Crash

I ran Furmark (stress test option) and temperatures kept rising until 85C and I then quit the application. - No Crash

 

[Endless8]

Hello _kain,
I see you did well with some research already. But let's start with few simple tests.
Download HWMonitor to check temps, prime95 to test CPU, UNrealHeaven, 3DMark or Furmark to test GPU and see what happens, no errors in MemTest is a good sign.
You can try removing drivers with DDU and install beta ones (if avaible), i had similar issue once and I fixed it like this, after the problem did not appeared with new drivers. You can also update DirectX.
Btw what is your PSU? and do you have the possibility to test your card in another PC or try another card (if you have one) in yours?
Cheers.

Hi Endless, thanks for the reply. I have OpenHardwareMonitor installed and on desktop temps seem fine (and the crash can happen when I'm just on the desktop and using a web browser) everything is about 30C except CPU which is about 35-37C.

I tried updating DirectX but already at the latest version. I will download prime95 and a GPU tester now and get back to you.

note: I cant remember the exact power of my PSU, but my system has been running fine for 1year+ (which is the last time I upgraded it when i bought my current PSU and GPU, and I made sure that I bought a PSU with more than enough power). Unfortunately I don't have another card/PC to test.

Ok, but I just want to make clear, more power PSU doesn't mean better, because cheap PSU can be 600W and deliver only 300W in real life + have no protection and unstable voltage which can be dangerous for your hardware, when a good quality PSU for example 500W will deliver 500W, have stable voltage and a lot of protections. Also cheap PSU get used over time quickly.

 

[_kain]

Hello _kain,
I see you did well with some research already. But let's start with few simple tests.
Download HWMonitor to check temps, prime95 to test CPU, UNrealHeaven, 3DMark or Furmark to test GPU and see what happens, no errors in MemTest is a good sign.
You can try removing drivers with DDU and install beta ones (if avaible), i had similar issue once and I fixed it like this, after the problem did not appeared with new drivers. You can also update DirectX.
Btw what is your PSU? and do you have the possibility to test your card in another PC or try another card (if you have one) in yours?
Cheers.

Hi Endless, thanks for the reply. I have OpenHardwareMonitor installed and on desktop temps seem fine (and the crash can happen when I'm just on the desktop and using a web browser) everything is about 30C except CPU which is about 35-37C.

I tried updating DirectX but already at the latest version. I will download prime95 and a GPU tester now and get back to you.

note: I cant remember the exact power of my PSU, but my system has been running fine for 1year+ (which is the last time I upgraded it when i bought my current PSU and GPU, and I made sure that I bought a PSU with more than enough power). Unfortunately I don't have another card/PC to test.

Ok, but I just want to make clear, more power PSU doesn't mean better, because cheap PSU can be 600W and deliver only 300W in real life + have no protection and unstable voltage which can be dangerous for your hardware, when a good quality PSU for example 500W will deliver 500W, have stable voltage and a mot more protections. Also cheap PSU get used over time quickly.

Just looked in the case - its an Antec 620W. My previous PSU failed and so when I bought this one I made sure to go with a recommended manufacturer. Also, I accidentally chose your reply as best solution, but the problem persists.

 

[Endless8]

Did you try Unreal heaven and 3DMark too? And what about drivers, did you uninstalled them, removed with DDU after and installed beta? Pretty strange, it passed GPU test.

 

[TDice]

Are you still having a problem????

 

[Endless8]

Btw, I have another idea, reset bios to default and do A MemTest with 1 stick of ram an test them 1 by 1.

 

[_kain]

Btw, I have another idea, reset bios to default and do A MemTest with 1 stick of ram an test them 1 by 1.

Just tried both unreal heaven (dx11) and 3d Mark11 and neither of them caused a crash. The issue is that these crashes happen at random (sometimes on desktop, sometimes after 10min of playing a game, sometimes I can use the computer for a couple days doing a bunch of different things without a crash).

As to the BIOS, I've not really played around with it before and I'm wary that I might completely mess it up. Also I've not updated it before so would that mean that its already at default?

 

[Endless8]

Btw, I have another idea, reset bios to default and do A MemTest with 1 stick of ram an test them 1 by 1.

Just tried both unreal heaven (dx11) and 3d Mark11 and neither of them caused a crash. The issue is that these crashes happen at random (sometimes on desktop, sometimes after 10min of playing a game, sometimes I can use the computer for a couple days doing a bunch of different things without a crash).

As to the BIOS, I've not really played around with it before and I'm wary that I might completely mess it up. Also I've not updated it before so would that mean that its already at default?

If you didn't touch it before that means its set to default.
Well, I would like to tell you to update the bios but I don't want you mess something.
Let's try the memtest like I said.